You are on page 1of 5

HR ANALYTICS AT SCALENEWORKS

BEHAVIORAL MODELING TO PREDICT RENEGE

1. Key Drivers:

I have considered below parameters as key drivers


1. Joining Bonus
2. Band Offered
3. Candidate relocate actual
4. DOJ Extended( Y/N)
5. Location
6. LOB offered
7. Status
8. Gender
9. Candidate Source

2. CODE RULES:
getwd()
scalene.df<-read.csv("Dataset_of_Scalene.csv",header=TRUE)
str(scalene.df)
scalene.df$DOJ.Extended = as.factor(scalene.df$DOJ.Extended)
scalene.df$Offered.band = as.factor(scalene.df$Offered.band )
scalene.df$Joining.Bonus = as.factor(scalene.df$Joining.Bonus )
scalene.df$Candidate.relocate.actual =
as.factor(scalene.df$Candidate.relocate.actual)
scalene.df$Gender = as.factor(scalene.df$Gender )
scalene.df$Candidate.Source = as.factor(scalene.df$Candidate.Source )
scalene.df$LOB = as.factor(scalene.df$LOB )
scalene.df$Location = as.factor(scalene.df$Location )
scalene.df$Status = as.factor(scalene.df$Status )
library(caTools)
set.seed(1)
split_indices=sample.split(scalene.df$Status,SplitRatio = 0.7)
train=scalene.df[split_indices,]
test=scalene.df[-split_indices,]
training_logit=glm(Status ~., data = scalene.df, family = binomial(link =
"logit"))
summary(training_logit)
library(rpart)
r.ctrl = rpart.control(minsplit=50, minbucket = 10, cp = 0, xval = 10)
## calling the rpart function to build the tree
m1 <- rpart(Status ~ ., data = train,control = r.ctrl, method = "class")
printcp(m1)
ptree = prune(m1, cp= 0.00677966 ,"CP")
prp(ptree,type = 5)
test$predict.class <- predict(ptree, test, type="class")
train$predict.class <- predict(ptree, train, type="class")
conf <- confusionMatrix(test$predict.class, test$Status)
conf
confusin_training <- confusionMatrix(train$predict.class, train$Status)
confusin_training
CONFUSION MATRIX:
3. INTERPRETATIONS:

 DOJ EXTENDED: There was a reduced renege on extended DOJs


 Candidate Source: Employee referrals noted lowest renege
 Joining Bonus: Had no effect on renege
 LOB: Lower renege has been noted in Infra
 If the notice period of the candidate is less than 38 days and is not
relocated, then the duration for the candidate to accept the offer is more
than 93 days and henceforth, the person is not likely to join the company.
 If the notice period of the candidate is greater than 38 days and is not
relocated, and if the HLT, MMS or INF then he is likely to join the
company.
 If the line of the business is AXO,BFS,CSM, EAS,ERS and ETS and the
duration to accept the order is less than 107 days and greater than 27 days
then the person is likely to join the office and if it is greater than 107 days
then the person is not likely to join the company.

You might also like